57 lines
980 B
Plaintext
57 lines
980 B
Plaintext
|
# Conception
|
||
|
|
|
||
|
|
## Intrinsic calibration
|
||
|
|
|
||
|
|
cf TP Simone Gasparini
|
||
|
|
|
||
|
|
## Extrinsic calibration
|
||
|
|
|
||
|
|
cf TP Simone Gasparini
|
||
|
|
|
||
|
|
## Skeletonization
|
||
|
|
|
||
|
|
- Input : vector string
|
||
|
|
- Ouptut : skeleton
|
||
|
|
|
||
|
|
### Points of interest detection
|
||
|
|
|
||
|
|
- Input : string : pictures ().jpg), (int : nb features?)
|
||
|
|
- Output : keypoints vector ( std::vector<KeyPoint>)
|
||
|
|
- function : we chose to use function SIFT (not SURF or Harris
|
||
|
|
|
||
|
|
- tests :
|
||
|
|
- with an image, verify the right number of points is detected
|
||
|
|
|
||
|
|
### Matching points
|
||
|
|
|
||
|
|
- Input : keypoints vectors, calibration
|
||
|
|
- Output : keypoints vector
|
||
|
|
- function :
|
||
|
|
- tests :
|
||
|
|
- use the function to match the points between a picture and herself
|
||
|
|
- use the function with different soft toys on the first and second picture
|
||
|
|
|
||
|
|
### Binary filter ?
|
||
|
|
|
||
|
|
- Input : binary filter, matched points
|
||
|
|
- Ouptut : filtered matched points
|
||
|
|
- tests :
|
||
|
|
-
|
||
|
|
|
||
|
|
### Camera filter
|
||
|
|
|
||
|
|
|
||
|
|
### Skeleton division
|
||
|
|
|
||
|
|
- Input :
|
||
|
|
- Output :
|
||
|
|
|
||
|
|
|
||
|
|
### Skeleton matching
|
||
|
|
|
||
|
|
- Input :
|
||
|
|
- Output :
|
||
|
|
|
||
|
|
|
||
|
|
|